giraph-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Maja Kabiljo (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (GIRAPH-543) Fix PageRankBenchmark and make WeightedPageRankBenchmark
Date Thu, 11 Apr 2013 00:21:17 GMT

    [ https://issues.apache.org/jira/browse/GIRAPH-543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13628485#comment-13628485
] 

Maja Kabiljo commented on GIRAPH-543:
-------------------------------------

It shoudln't be GC because I conducted experiment with much more memory than needed. The data
is smaller this way, and the total size of requests too. When multithreaded computation is
used, we spend some time waiting on open requests to finish in the end of computation, and
I suppose that's why we get that big speedup. Although even with single-thread I got like
70s->50s, since we do have less things to serialize/deserialize and we compute only with
floats.
                
> Fix PageRankBenchmark and make WeightedPageRankBenchmark
> --------------------------------------------------------
>
>                 Key: GIRAPH-543
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-543
>             Project: Giraph
>          Issue Type: Bug
>            Reporter: Alessandro Presta
>            Assignee: Maja Kabiljo
>         Attachments: GIRAPH-543.patch
>
>
> PageRankBenchmark uses PseudoRandomInputFormat, which produces edges with double values.
> Since edge values are meaningless for the algorithm, we should change it to use null
values to have a better evaluation of PageRank performance.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message